Response to Reviewers Aged streams: Time lags of nitrate, chloride and tritium assessed by Dynamic Groundwater Flow Tracking

to discuss t he spatial and temporal effects of land-use and groundwater flow paths. We will subsequently discuss the case of the study area and take another look at the observed breakthroughs and time lags of the different solutes, varying different scenario and evaluating model behaviour and outputs. Lastly, we will discuss the implications of our findings for catchment management.” abstract include more the aspects and Revised text: “ This paper is one of the first to calculate in-stream concentrations of tritium, chloride and nitrate using dynamic groundwater travel time distributions (TTDs) derived from a distributed, transient 3D groundwater flow model using forward particle tracking. ” “ We tested our approach in the Springendalse Beek catchment, a lowland stream in the east of the Netherlands, ” “ By making the connection between dynamic groundwater travel time distributions and in-stream concentration measurements, we provide a method for validating the travel time approach and make the step towards application in water quality modelling and management. Important conclusion for management is that although the effect of mitigation measures can generally be expected instantly, based on an almost exponential steady state distribution of travel times in a catchment, our model also showed that the spatial distribution of agriculture in catchments influence the response of stream concentrations to mitigation actions. ”

In the discussion we will first focus on the effect of processes on the breakthrough of solutes, based on the exploration of the model behaviour under different scenarios. We will clarify the concept of time lags we proposed and will evaluate how unsaturated zone processes, advective saturated flow, the spatial source zones and temporal patterns of solute inputs and transformation processes such as denitrification all will influence these time lags in the receiving water body. Using our dynamic TTD model, we will then introduce 'contributing areas' to discuss the spatial and temporal effects of land-use and groundwater flow paths. We will subsequently discuss the case of the study area and take another look at the observed breakthroughs and time lags of the different solutes, varying different scenario and evaluating model behaviour and outputs. Lastly, we will discuss the implications of our findings for catchment management." Also the model lacks measures to evaluate its performance/fit.

RESPONSE:
The groundwater model itself was calibrated and validated using groundwater heads and discharges (discharge NSE 0.65), as discussed in Kaandorp et al., 2018 WRR. The model was not further calibrated as it is only used as a tool to investigate it's use for the particle tracking / Travel Time method. In the analysis done in the current paper the measured chloride, nitrate and tritium function as an added evaluation of the transient TTD model.
Likewise the sensitivity analysis does not include a comprehensive evaluation of the parameter importance.

RESPONSE:
Agreed. The term 'sensitivity analysis' was a bit misleading for what we tried to do. We have changed this term throughout the manuscript to an 'exploration of the model behavior under different scenarios', to better represent the actual method and goal. The effect of the different parameters is discussed in section 4.1 (The effect of processes on modelled breakthroughs) and we will add a more comprehensive evaluation to this paragraph to indicate the importance of the different parameters and processes.

RESPONSE:
Agreed. We will extent the discussion section with further discussion on the responsible hydrological processes that play a role. We will revise the conclusions to better represent the innovative parts of the manuscript. Additions to the text include: Revised text: "By making this connection between dynamic groundwater travel time distributions and in-stream concentration measurements, were able to show that the seasonal and long-term fluctuations of in-stream solute concentrations were mainly caused by the dynamic contribution of different groundwater flow paths." "We show that for most catchments, with a (close to) exponential TTD, a direct reaction of stream nitrate concentrations can be expected to reductions in the nitrogen inputs. Furthermore, we found that the breakthrough patterns of diffuse pollutants are largely governed by the spatial distribution (distance) of agricultural fields and the reactivity of the subsurface, which can both be tackled by creating riparian buffer zones." The authors manage to write comprehensively and with a good overall structure. Just every now and then they should try to be more consistent when using short versions (Cl/NO3) or the words "chloride"/"nitrate".

RESPONSE:
Thanks for the compliment regarding the comprehensiveness and structure. We now use 'chloride' and 'nitrate' consistently.

Revised text: "chloride" and "nitrate"
When reading the title of this manuscript I was hoping for more novelty in the results and discussion sections. Yes, the study is well-written and has a good structure, but somehow I feel that the authors do not go much beyond reporting the results. There is neither a lot of analysis nor discussion on which controls are important and specifically why they are more or most important. I like the modeling approach although it bothers me that it is purely a groundwater model without unsaturated zone processes. However, I believe that the authors can manage to add more analysis and discussion to merit and justify a publication in HESS.

RESPONSE:
We agree with the reviewer. Especially the discussion and conclusions do not capture the innovative parts of the manuscript in the right way. Following the reviewers suggestions, we will critically go through the discussion section and add more discussion on the processes and controls. It is important to mention here that this paper is one of the first to couple measurement data of water quality and isotopes to Travel Time Distributions. Even though many recent papers suggest that this method holds a lot of potential, the papers which actually attempt to do so are absent or at least very scarce.
In this modelling study we focus on the groundwater system and indeed highly simplify the unsaturated zone. The unsaturated zone is included in the method in two ways: 1) the groundwater model is coupled with an unsatured zone model (MetaSWAP, see also Kaandorp et al., 2018 WRR) to provide a realistic groundwater recharge based on e.g. land-use; 2) the input curves include part of the unsaturated zone processes: before 2000 by using a nitrate transformation factor of 0.85 and after 2000 by using the concentrations of the shallow groundwater to construct the input curves. Thus, we take into account many of the unsaturated zone processes, just not the delay.
Furthermore, the relevance of the unsaturated zone also depends on the research area and the temporal scale of interest. In this case, the research area is a lowland groundwater-driven catchment with shallow groundwater levels and the focus is on seasonal changes in flow paths and solute concentrations in the stream. This can be well simulated with a simple representation of the unsaturated zone. In a more hilly catchment, or if we would want to simulate the concentration response to individual events, a more detailed representation of unsaturated zone hydrology would be required.
We will also highlight these more in the manuscript.

Specific comments:
Abstract: Try to underline the innovative parts of your study more. Think about more precise results (like TT results, R2) Isn't it important to at least name the investigation area?

RESPONSE:
We agree and we thank the reviewer for these suggestions and will rewrite the abstract to include more the innovative aspects and precise results. We added for instance: Revised text: "This paper is one of the first to calculate in-stream concentrations of tritium, chloride and nitrate using dynamic groundwater travel time distributions (TTDs) derived from a distributed, transient 3D groundwater flow model using forward particle tracking." "We tested our approach in the Springendalse Beek catchment, a lowland stream in the east of the Netherlands," "By making the connection between dynamic groundwater travel time distributions and in-stream concentration measurements, we provide a method for validating the travel time approach and make the step towards application in water quality modelling and management. Important conclusion for management is that although the effect of mitigation measures can generally be expected instantly, based on an almost exponential steady state distribution of travel times in a catchment, our model also showed that the spatial distribution of agriculture in catchments influence the response of stream concentrations to mitigation actions." p.2, 33: "through" instead of "though"?

RESPONSE: Agreed and changed accordingly
Revised text: "through" p.2, 38: how do you mean "land-use" in this context?

RESPONSE:
We agree that this is not entirely clear. We mean for instance whether or not manure/fertilizers are used, we made this more clear in the text. Revised text: "depends on these groundwater flow paths and travel times, as well as on land-use (for instance use of manure and fertilizer)." p.2, 41: "Land-use determines the timing and quantity of nitrate input" would be more general RESPONSE: Agreed and changed accordingly.
Revised text: "Land-use determines the timing and quantity of nitrate input as well as its spatial distribution" p.2, 42: ".depending on the source and timing of infiltration" RESPONSE: Agreed and changed accordingly.
Revised text: "depending on the source and timing of infiltration," p.2, 52: What time is meant by "historical"?

RESPONSE:
We agree that this is not clear. We changed the text into: Revised text: "Therefore, we combine in this follow-up paper these dynamic travel time distributions with input curves for tritium, chloride and nitrate to reconstruct the concentrations in a Dutch stream from 1970 until present, while including the dynamic nature of catchments both in time and space." p.3, 74: If 20 % of the river are not baseflow, is this not an important NO3 source due to surface runoff with high solute concentrations? p.5, 140: What are the distances of the two stations to the corresponding measurement station? RESPONSE: For both stations the distance is around 80 to 90 km. We will add this information to the manuscript.
p.5, 148: "following Meinardi (1984)" Revised text: "following Meinardi (1994)" p.5, 154: How can you exclude land-use change? RESPONSE: Land-use change will obviously affect the input of nitrate and we did not exclude this. Land use change is partly incorporated in the input curves. Furthermore, a large agricultural field in the upstream part of the catchment has been in use only since approximately 1985 and an agricultural field of approximately 7.5 ha was converted to natural vegetation in 1998. This was also considered when combining the TTDs and input curves. What was not considered is that small parts of the catchment changed from nature in agriculture and vice versa. This does not have a large effect on the downstream surface water concentrations. We simply do not have enough data to take small changes into account: -we do not have knowledge on the high detail and long term land-use changes (e.g. specific crops), and -farmer specific differences will affect the result (using maximum amount of fertilizer or not, etc). We do discuss the effects of this assumption later in the Discussion section.
Revised text: "Land use change is partly incorporated in the input curves. Furthermore, a large agricultural field in the upstream part of the catchment has been in use only since approximately 1985 and an agricultural field of approximately 7.5 ha was converted to natural vegetation in 1998. This was also considered when combining the TTDs and input curves. What was not considered is that small parts of the catchment changed from nature in agriculture and vice versa. This does not have a large effect on the downstream surface water concentrations." p.5, 156: How can you assume NO3 to be constant over time although atm. N-deposition changed strongly within the last decades? p.8, 236: It is also assumed as a first-order decay (e.g. van Meter et al., 2017) in literature. Why do you assume zero-order? RESPONSE: Much literature exists on this matter and we considered using a first-order decay equation as well. However for simplicity, we decided to only use the zero-order decay and the complete decay at specific depth here.
p.8, 240: Do you have information about the denitrification potential assuming it as a finite process? RESPONSE: In this scenario we assume the reactivity potential of the subsurface to be an infinite process. This complete decay at depth is in correspondence with literature from the Netherlands (Zhang et al., 2013;Visser et al., 2009) and Denmark (Jessen et al., 2017;Koch et al., 2019). p.8, 249: Figure 3d as well? In the following sentence, I propose to add "Figure 3d". RESPONSE: Agreed, we changed the text according to the reviewers suggestion.
p.8, 252: Highest contribution in winter although highest N inputs can be assumed in spring and are washed out until winter? And how can you see seasonal NO3 patterns assuming constant NO3 input over the year? So I think "agricultural contribution" (assumed as constant) is not the appropriate wording, perhaps NO3 transport/mobilization?

RESPONSE:
The text here and Figure 3c and d, refer to the origin of water, i.e. infiltrated on agricultural fields. We will make this clearer in the text. p.9, 268 I cannot associate the word "later" with the mentioned years.

RESPONSE: Agreed and changed accordingly.
Revised text: "which is about 5 to 10 years after the peak in the input (Figure 2). " p.9, 273: What is a reasonable fit? Do you have a measure for this, like R2? RESPONSE: We will try to give it a measure to describe the fit.
p.10, 297: What is meant by "did not completely fit"? Think about using a measure to show the actual fit. RESPONSE: See previous comment. p.9, 284-286: Would not denitrification cause an earlier peak/decay than Cl instead of the surprisingly later peak. Is the changing N-Cl ratio not excluded by the input data of the farmer (chapter 2.5)? Do you have any hints for potential point sources? p.9, 286: What point sources are realistic in a non-urban area? RESPONSE: We will more closely analyze and discuss the effects of denitrification and the difference in the breakthrough of different solutes. We will also look into potential point sources.
p.10, 295: Can you adjust the headings in this chapter or in the according method section so that it is easier to assign? RESPONSE: Agreed. We will make this more consistent.
p.11,321: Think about adding ( Figure 6, "red line") and later on (. . ., "green line"). Than it is easier to match. But this is probably a matter of taste.